White Paper. Who s Getting Paid During the Subprime Crisis?

Size: px

Start display at page:

Download "White Paper. Who s Getting Paid During the Subprime Crisis?"

Janel Flora Manning
6 years ago
Views:

1 > White Paper Who s Getting Paid During the Subprime Crisis? Jennifer Christensen, Senior Consultant Yara Rogers-Silva, Consulting Statistician III May 2008

2 Table of Contents Executive Summary Methodology Observation Period Equifax Consumer Modeling Attributes Definitions Statistical Testing Study Findings Overall Results Payment Behavior by State Amount Payment Behavior by Risk Score Age of Oldest Trade Recent Delinquency Revolving High Credit Number of Inquiries Within 12 Months Utilization Conclusions

3 Executive Summary delinquencies have been increasing at a rapid pace since first quarter Much of this delinquency is driven by higher monthly mortgage payments caused by ARMs (adjustable rate mortgages) that have entered their reset period. Another contributor is the record amount of equity withdrawal, which has led to high debt service burdens and more financial pressure on consumers. If consumers are unable to make all of their monthly payments, they must choose which obligations they will pay each month, and which they will either pay late, partially or not at all. In some instances, consumers will be forced to choose between their bankcard, auto and mortgage. Traditional industry consensus is that consumers typically pay their bills according to a certain hierarchy that puts mortgage first, followed by auto, card and utility bills. Recently, however, there has been speculation about whether or not there has been a change in this hierarchy such that consumers are prioritizing their bankcard and auto loan payments over their mortgage payment. The objective of this study is to quantify how prevalent this reprioritization is and to provide insight into the type of consumers who exhibit this behavior. Traditionally, consumers typically paid mortgages first, then auto loans, credit card and utility bills. Observations This study found that 1.2 million consumers opened a mortgage in July 2005 and that 3.4% of them had a 60-day mortgage delinquency over the next 24 months. Of those that had a mortgage delinquency in the 24-month performance period from July 2005 to July 2007, 38% kept their bankcards clean and 62% kept their auto loans clean. As a means of comparison, Equifax conducted a similar study on mortgages opened in June That study found that of consumers who had a mortgage delinquency in the 24-month performance period from June 2002 to June 2004, 26% kept their bankcards clean and 59% kept their auto loans clean. This shows that the priority consumers give to auto loan payments has not changed much over time; however, there is a marked difference in bankcard payment priority. 2

4 This study used a subset of highly predictive attributes and scores, calculated as of July 2005, to create a profile of consumers who keep their auto and bankcard loans current over the performance window, while struggling to meet their mortgage payments. The profile shows that these consumers have: More established credit histories. More trade lines. Better risk scores. Lower utilization of revolving lines. Fewer recent inquiries. Fewer historical delinquencies on their credit histories. Methodology Observation Period The study included all mortgages opened in the U.S. during July 2005, with 24 months of performance through July This mortgage vintage was selected because it is known to have poor performance relative to other mortgage vintages. It also reflects recent mortgage lending practices, such as the prevalence of subprime and adjustable rate alternatives. Equifax Consumer Modeling Attributes This study makes use of Equifax Consumer Modeling Attributes (CMA Plus). The CMA Plus attributes are a comprehensive aggregation of consumer credit behaviors, giving Equifax customers a 360-degree view of the consumer credit file. By evaluating all aspects of the consumer credit file from credit history and type of accounts to delinquency patterns, public records and consumer-generated inquiries, the CMA Plus utilizes the full Equifax national consumer credit database to assist in predicting the future behaviors of consumers. The CMA Plus is a powerful tool that gives insight into consumer credit behavior patterns to drive decision making processes across the account life cycle. CMA Plus attributes are used in this study for attribute-level analysis of specific segments of the population. 3

5 Definitions Term Definition Population s opened in the U.S. during July 2005 Window Delinquency Calculation Auto Delinquency Calculation Delinquency Calculation Figure 1: Definitions Statistical Testing 24 months from July July days past due over the performance window on any mortgage from the July 2005 vintage. that is less than 60 days past due was considered clean. 60+ days past due over the performance window on any auto trade in the credit file with an open status as of July 2005 or opened within 18 months after the mortgage open date. Delinquency less than 60 days past due was considered clean. 60+ days past due over the performance window on any bankcard in the credit file with an open status as of July 2005 or opened within 18 months after the mortgage open date. Delinquency less than 60 days past due was considered clean. The non-parametric Wilcoxon Signed rank test was used to confirm that the differences in the means of the variables shown in the paper are indeed statistically significant. Study Findings Overall Results Figure 2 below shows that consumers from the June 2002 mortgage vintage who had clean mortgage performance were more likely to have an auto or bankcard delinquency than their counterparts from the July 2005 vintage. This suggests that, in the past, consumers may have given higher priority to mortgage payments and lower priority to auto and bankcard payments, versus what we see today. Vintage Clean Clean Auto Auto Clean % 7.5% 97.0% 3.0% % 5.8% 97.5% 2.5% % 73.6% 59.4% 40.6% % 62.1% 62.0% 38.0% Figure 2: v. Auto Loan and 4

6 Figure 3 shows that consumers in the June 2002 mortgage vintage with delinquent auto performance were more likely to have bankcard delinquency than consumers in the 2005 vintage (67.5% and 62.1%). This suggests that, in recent times, bankcards are receiving higher priority from consumers than in the past. Conversely, those in the 2002 vintage with delinquent auto performance were more likely to have mortgage delinquency than those in the 2005 vintage (33.4% vs. 36.2%), suggesting that mortgage payments have lower priority now than in the past. payment has become more important, while auto loan payment priority has changed relatively little. In both 2002 and 2005, consumers with a delinquent auto tradeline were more likely than not to have bankcard delinquency (67.5% and 62.1%). Furthermore, consumers in both vintages with bankcard delinquency were less likely to have delinquent auto trades (24.6% vs. 25.0%). This suggests that consumers place a higher priority on auto loan payments than bankcard payments. Results also suggest that auto loan payment priority has not changed much over time, while bankcard payment priority has become more important. Clean Auto Auto Vintage Clean Clean % 7.8% 97.8% 2.2% % 6.1% 97.7% 2.3% % 67.5% 66.6% 33.4% % 62.1% 63.8% 36.2% Figure 3: Auto Loan v. and Overall, consumers continue to place the highest priority on mortgage payments. Of the consumers with delinquent auto payments in the 2002 vintage, 66.6% kept their mortgage clean compared to 63.8% in the 2005 vintage. Findings were similar for consumers with delinquent bankcards in the 2002 and 2005 vintages, where 78.7% and 76.7% kept mortgages clean. 5

7 Clean Auto Vintage Clean Auto Clean % 1.3% 99.2% 0.8% % 1.3% 98.9% 1.1% % 24.6% 78.7% 21.3% % 25.0% 76.7% 23.3% Figure 4: v. Auto Loan and With no equity in their homes, consumers are more likely to forfeit mortgage payments in order to keep autos and bankcards. Payment Behavior By State This study found that states that had the highest home prices as of December 2005 (the period with peak home prices), followed by large price declines during the study's performance window, had the highest percentage of consumers who kept their auto and bankcards clean, while letting their mortgage go delinquent. In these states, many consumers stretched to purchase expensive homes, assuming that home prices would continue to rise. Unfortunately, median home price values began to drop starting in the first quarter of 2006, causing consumers to become "upside down" on their mortgages (the condition when the mortgage balance exceeds the value of the house). With no equity in their homes, many of these consumers appeared willing to walk away from their mortgage in order to keep their auto and access to their credit cards. Figure 5 below shows auto and bankcard delinquency for consumers who had a mortgage delinquency over the 24-month performance window in Arizona, California, Florida, Massachusetts, Maryland, and Nevada (states with high home price peaks and subsequent falling home prices), compared to the rest of the country. Population: Consumers with Delinquency State AZ, CA, FL, MA, MD, NV Clean s s Clean Auto Trades Auto Trades 43% 57% 67% 33% All Other States 36% 64% 60% 40% Total 38% 62% 62% 38% Figure 5: and Auto Loan by State 6

8 Amount Figure 6 shows that the mortgage loan amount may influence consumer behavior. Consumers with clean bankcard and auto loans have higher average mortgage amounts than those with auto and bankcard delinquencies. Consumers with higher mortgage amounts are likely to have higher mortgage payments and may find themselves in an adverse financial position due to ARM resets or other life events. In many cases, they can simply no longer afford their mortgage payment and need to walk away from their home in order to retain their auto and access to credit via their credit cards. Amount Auto Amount $185,466 $177,715 Clean $217,737 Clean $198,846 Clean $178,322 Clean $172,206 Clean Clean $214,588 Clean Clean $210,927 Figure 6: Amount by, and Auto Loan Payment Behavior by Risk Score Figures 7 and 8 show payment behavior stratified by Beacon 5.0 score quartiles at the loan origination time (July 2005). Beacon 5.0 is a generic risk score that Equifax co-developed with Fair Isaac Corporation. It predicts the likelihood that a consumer will become a serious credit risk within 24 months from scoring. In the three lowest scoring quartiles, consumers were the most likely to have auto and bankcard delinquencies. This population is a high credit risk population and consumers in this score range are likely to have displayed poor credit payment behavior in the past. They are also likely to continue this behavior in the future. Consumers who fall into the fourth quartile, which represents scores 663 and higher for the population with bankcards and 653 and higher for the population with auto trades, are the most likely to continue to pay their other financial obligations. Even in the event of mortgage default, this population may be an area of opportunity, especially for auto lenders, as 72 % of consumers in this quartile will keep their auto loan clean even in the event of a mortgage default. Further analysis also demonstrated, unsurprisingly, that those consumers who kept current on their bankcard and auto tradelines had higher risk scores at observation, compared with those that went delinquent. 7

9 Figure 7: by BEACON 5.0 Figure 8: Auto Loan by BEACON 5.0 8

10 Consumers who maintain both bankcard and auto loans, even if they default on their mortgages, typically have longer credit histories than those who do not. Age of Oldest Trade Figure 9 displays the attributes Age of Oldest Trade, Age of Oldest Trade, and Age of Oldest Auto Trade. These attributes are a measure of length of credit history. Both trade age attributes show that consumers who maintain their bankcards and auto loans even if they go delinquent on their mortgage have longer credit histories on average than those who do not. Consumers with more established credit histories may recognize the benefits of maintaining a good credit history, such as more credit options and more favorable interest rates. Avg. Age Oldest Trade Avg. Age Oldest Trade Auto Avg. Age Oldest Trade Avg. Age Oldest Trade Clean Clean Clean Clean Clean Clean Figure 9: Age of Oldest Trade (months) by, and Auto Loan Recent Delinquency Figure 10 shows that recent delinquency is correlated with auto and bankcard payment behavior. As expected, consumers with fewer historical delinquencies are more likely to maintain their auto loans and bankcards going forward. Consumers with previous delinquency may have limited capacity or willingness to pay their monthly financial obligations as is reflected in the following tables. 9

11 Number 30-Day Number 30-Day Past Dues Number 30-Day Past Dues Number 30-Day Past Within 24 Within 24 Past Dues Dues Within 24 Mos. Mos. Accts. Auto Mos. Auto Accts. Within 24 Mos. Auto Accts Clean Clean Clean Clean Clean Clean Figure 10: Number 30 Days Past Due Within 24 Months by, and Auto Loan As expected, creditworthy customers are more likely to keep their auto loans and bankcards current should they become delinquent on their mortgages. Revolving High Credit Figure 11 below shows that the revolving high credit amount is correlated with auto and bankcard payment behavior. The available credit amount is a proxy for how credit worthy the customer is. As expected, more creditworthy customers are more likely to keep their auto loan and bankcards current in the event that they become delinquent on their mortgages. Presence of 60 DPD High Credit Open Revolving Trades Auto High Credit Open Revolving Trades $22,128 $18,925 Clean $29,612 Clean $25,698 Clean $31,592 $25,509 Clean Clean $59,177 Clean $54,587 Figure 11: Revolving High Credit by, and Auto Loan Number of Inquiries Within 12 Months Figure 12 shows that the number of recent inquiries is correlated with auto loan and bankcard payment behavior. A flurry of recent credit inquiries suggests that the consumer is in financial trouble and is seeking additional credit. Recent credit inquiries may also indicate that additional lines of credit may have been issued to a consumer, but have not yet appeared on the consumer's credit report. This is a telltale sign that the consumer's debt-to-income ratio may be higher than anticipated. 10

12 Number of Inquiries Within 12 Months Auto Number of Inquiries Within 12 Months Clean 6.1 Clean 6.5 Clean Clean Clean 3.9 Clean 4.2 Figure 12: Number of Inquiries in 12 Months by, and Auto Loan Utilization Figures 13 and 14 show that consumers with lower revolving credit utilization are more likely to maintain clean bankcard and auto loans in the event of mortgage delinquency. High revolving utilization may indicate that a consumer is unable to afford their monthly credit obligations and may be using credit cards to pay their bills. This trend is more observable in bankcard payment behavior than in auto payment behavior. Figure 13: by Utilization 11

13 Figure 14: Auto Loan by Utilization Conclusions The purpose of this research paper was to verify an emerging segment of consumers who maintain their auto and bankcard loans, while letting their mortgage go delinquent. The analysis clearly demonstrated that there are, in fact, consumers who display this type of payment behavior. The prevalence of this behavior was somewhat surprising. In the event of mortgage delinquency, consumers are more likely to continue making their auto payment than their bankcard payments. Of consumers that fall behind on mortgage payments, 62% kept their auto loan current, while 38% kept their bankcards current. This is understandable since consumers rely on their autos to maintain their source of income; you can't drive your bankcard to work. The business implication is that traditional assumptions about consumer payment behavior may not bear out over irregular economic conditions, such as the recent mortgage industry plight. Of course, this general business implication may not apply equally to all portfolios. This analysis does demonstrate, however, that credit bureau data can help identify consumers who are likely to exhibit payment behavior that is contrary to what is accepted as the traditional payment hierarchy. Armed with this information, 12

14 financial institutions may be better able to quantify the spillover effect from mortgage to auto and bankcard, which may lead to different portfolio management strategies and likely less dramatic policy shifts. About the Authors: Jennifer Christensen, Senior Consultant Jennifer Christensen has been with Equifax for four years and is responsible for client relationships, including evaluating client strategies and recommending appropriate, profitability driven solutions. Jennifer has expertise in collections, and line management strategy development and implementation. Prior to joining Equifax, Jennifer was Senior Credit Risk Manager at Metris, where she managed the team that was responsible for developing, evaluating and implementing collections and line management strategies. This team was integral in preparing for federal financial audits and for responding to their findings. Jennifer brings over 16 years of financial industry experience, including three years as a business analyst at Trans Union and three years at First Card, where she initially served as preapproved process manager and then as credit analyst supporting the account management team. Yara Rogers-Silva, Statistical Consultant III Yara Rogers-Silva joined Equifax in January of She has worked on a variety of customer engagements representing various industries, as well as developing risk models for Equifax Latin America operations. Previously, Yara was Vice President of statistical modeling at SunTrust Banks, where she was responsible for the Statistical Modeling of all marketing programs for deposit products. Prior to SunTrust, Yara worked as a Senior Operations Research Consultant at Delta Airlines developing predictive models for the maintenance division and managing revenue. Yara holds an M.S. in Operations Research and Decision Sciences from Rensselaer Polytechnic Institute and a BS in Applied Mathematics from the State University of Campinas, Brazil. 13

15 Equifax is a registered trademark of Equifax Inc. Copyright 2008, Equifax Inc., Atlanta, Georgia. All rights reserved. Do not copy or reproduce any part of this document without express written authorization from Equifax. EFS-769-ADV 5/08

Are today s market pressures reshaping credit risk?

Are today s market pressures reshaping credit risk? New study explores FICO Score trends in dynamic times and how lenders can respond Number 3 May 2008 In turbulent economic times, financial services firms